a option to disable the output of unmapped query when --fmt 2

[wangyugui]

Hi.

Is there a option to disable the output of unmapped query when --fmt 2?

We need to output all mapped query only.

And a format like blast format 6 will be good too.

[wangyugui]

currently I am using

cmscan --cpu 2 --rfam --cut_ga --nohmmonly --tblout mrum-genome.tblout --noali --notextw --oskip --fmt 2 --clanin /usr/bio-ref/rfam-12.3/Rfam.clanin /usr/bio-ref/rfam-12.3/Rfam.cm -

[nawrockie]

Can you please provide an example of what you mean by 'unmapped query'?

It is not possible to output 'blast format 6' exactly, but the tabular output file (mrum-genome.tblout) is similar. That tabular file can be easily manipulated using grep and awk, like this:

cat mrum-genome.tblout | grep -v ^# | awk '{ printf("%s %s %s %s %s\n", $2, $4, $10, $11, $18); }'

which will print only the model name, sequence name, sequence start position, sequence stop position and E-value for all hits.

[nawrockie]

The tabular output file format is described on page 60 of the Infernal v1.1.2 user guide, available from: eddylab.org/infernal/.

[wangyugui]

an example of 'unmapped query':

Query:       TRINITY_DN94_c0_g1_i1  [L=455]
Description: top_iso:1 pct_iso_expr=100 pct_dom_iso_expr=100
Hit scores:
 rank     E-value  score  bias  modelname  start    end   mdl trunc   gc  description
 ----   --------- ------ -----  --------- ------ ------   --- ----- ----  -----------

   [No hits detected that satisfy reporting thresholds]

Internal CM pipeline statistics summary:
----------------------------------------
Query sequence(s):                                               1  (910 residues searched)
Query sequences re-searched for truncated hits:                  1  (706.3 residues re-searched, avg per model)
Target model(s):                                              2687  (346722 consensus positions)
Windows   passing  local HMM SSV           filter:            1920  (0.0669); expected (0.06)
Windows   passing  local HMM Viterbi       filter:             566  (0.01923); expected (0.02)
Windows   passing  local HMM Viterbi  bias filter:             300  (0.008982); expected (0.02)
Windows   passing  local HMM Forward       filter:              39  (0.00145); expected (0.0002)
Windows   passing  local HMM Forward  bias filter:              26  (0.001074); expected (0.0002)
Windows   passing glocal HMM Forward       filter:              11  (0.0004396); expected (0.0002)
Windows   passing glocal HMM Forward  bias filter:              11  (0.0004396); expected (0.0002)
Envelopes passing glocal HMM envelope defn filter:              11  (0.0003051); expected (0.0002)
Envelopes passing  local CM  CYK           filter:               3  (6.516e-05); expected (0.0001)
Total CM hits reported:                                          0  (0); includes 0 truncated hit(s)

Thanks for your info.

mrum-genome.tblout is what I wanted. I'm sorry that I though this 'tbl' as 'tblast' not as 'tabular '

[wangyugui]

The final process maybe useful for others.

cat seq.fa |
parallel --halt soon,fail=1 --gnu --no-notice --keep-order --pipe -j 12 --block 100k --recstart '>' \
    cmscan --cpu 2 --rfam --cut_ga --nohmmonly \
    --fmt 2 --clanin Rfam.clanin --tblout /dev/stderr \
    Rfam.cm -  >rfam.scan.txt 2>rfam.scan.tbl.txt